Forgepoint 籌集 $9.8M 種子輪融資，將 AI 引入製造業報價流程

A manufacturer can have the equipment, materials and expertise to make a part—and still lose the job while its quote waits in an inbox. Forgepoint is building an AI inside sales representative called Frank to shorten that gap between a customer’s request and a price the shop is ready to stand behind.

The New York company has raised a $9.8 million seed round led by Primary Venture Partners and a16z speedrun, with participation from angel investors associated with OpenAI, Brex, Vercel and SoFi. The funding will support expansion to more manufacturers, deeper integrations with existing business systems and hiring.

Its starting point is the request-for-quotation, or RFQ: a package of emails, drawings, specifications and commercial requirements that must become an accurate offer. Forgepoint says Frank handles the administrative work across that process while engineers, estimators and salespeople retain technical and commercial decisions.

The Bottleneck Before the Production Line

Factory automation often brings robots and machine tools to mind. Quoting exposes a different constraint. A shop cannot win work it never gets around to pricing, and its most experienced people may spend much of their day gathering information rather than evaluating the job.

An incoming RFQ can require a succession of small tasks: identify the correct drawing revision, confirm quantities, ask for missing specifications, find comparable orders, estimate costs, obtain engineering approval and prepare the customer’s document. Information then has to be entered into enterprise resource planning and customer relationship management systems, followed by further correspondence.

Each handoff introduces waiting time. Automating only the final document leaves much of that delay intact. Forgepoint’s approach is to connect intake, preparation, approval and follow-up around the manufacturer’s existing processes.

This sits within a broader shift toward applying generative AI across the manufacturing lifecycle, including the commercial and engineering work that surrounds physical production. The immediate opportunity here is administrative capacity: helping an existing team consider more potential jobs without increasing manual coordination at the same rate.

How Frank Moves an RFQ Toward an Approved Quote

According to the announcement, Frank receives an email address on the manufacturer’s domain and is configured around its operating procedures, pricing rules and approval requirements. When a request arrives, it reads the message and attachments, identifies gaps and contacts the buyer for clarification.

It then prepares pricing using the shop’s own cost model, brings forward similar historical jobs and routes the draft to the appropriate reviewer. After approval, it produces the final quote in the company’s template, sends it and manages follow-up. Emails, attachments and the reasoning behind decisions are preserved as a record of the work.

The company’s product documentation adds detail about document extraction and review: Frank is designed to pull dimensions, materials and tolerances from drawings and CAD files, apply labor rates and margins, and let staff edit a draft before approving it. It also describes configurable follow-up and pipeline reporting.

The important control point is approval. A quote represents a commercial commitment, and small mistakes in a tolerance, material requirement or cost assumption can change whether a job is profitable or feasible. Frank’s described role is to prepare that decision and carry out the surrounding workflow, with human review before a quote is sent.

Forgepoint has not disclosed the underlying model architecture in the announcement. Its differentiation rests on how AI is connected to manufacturing information and actions, rather than on a newly announced foundation model.

One Workflow, Different Shop Economics

Quoting is not a uniform calculation across manufacturing. A metal service center may need to match material lists and cut lengths against its catalog, check stock and apply customer-specific pricing. A foundry may first need an engineer to interpret a part file and resolve missing drawing requirements.

Forgepoint says it configures Frank to those differences, with applications spanning injection molding, fabrication, stamping, machining and other manufacturing processes. That customization is central to the product: the same request can imply different operations, costs and risks depending on the shop receiving it.

Historical records can help, but they also require judgment. A similar part quoted last year may have been made under different material prices, labor assumptions or customer terms. The usefulness of a remembered job depends on whether the system exposes the relevant context and lets reviewers correct assumptions.

Early Results and the Institutional Memory Problem

Forgepoint says it reached seven-figure annual recurring revenue within six months of Frank’s launch, with hundreds of engineers, estimators and sales representatives using the product weekly. Across customers, it reports an average 30% increase in monthly quote throughput and a reduction of more than 55% in time to quote.

Those are company-reported results. The announcement does not provide the customer sample, measurement period or independent validation needed to assess how widely they generalize. More quotes and faster responses also do not, on their own, establish better win rates or higher margins. They measure progress on a specific operational bottleneck.

Another part of the proposition is preserving knowledge. Pricing exceptions, customer preferences and lessons from earlier jobs often remain in experienced employees’ inboxes or memories. Capturing the correspondence and decisions behind a quote could make that context easier to recover when a colleague takes over or a veteran estimator retires.

Manufacturing Experience Meets Enterprise Software

Co-founder and CEO Jack Strauss previously worked at Northrop Grumman, in Bain & Company’s Advanced Manufacturing practice and at aerospace systems integrator Xwing. Co-founder and CTO Alex Pareto previously led engineering for Brex’s Spend Management organization.

The founders spent months visiting Midwestern plants in early 2025, shadowing employees to understand their work. The product reflects that focus on the surrounding process: existing emails, documents and pricing logic are the starting point for automation.

Forgepoint plans to extend its approach into other manufacturing back-office functions over time. For now, the seed round backs a narrower test with tangible measures: whether Frank can help shops prepare accurate quotes sooner, keep approvals visible and give skilled employees more time for the decisions that determine which work is worth taking on.